|Institution:||University of Oklahoma|
|Keywords:||Meteorology; Hydrology; Flooding|
|Full text PDF:||http://hdl.handle.net/11244/44865|
This study will describe the MRMS reanalysis precipitation dataset created for the time period from 2001 to 2011. This high resolution 1-km2 5-minute dataset is ideal for simulating flash floods with a distributed hydrologic model. The Ensemble Framework For Flash Flood Forecasting (EF5) is created for the purpose of exploiting this high resolution precipitation information by conducting simulations with multi water balance models. The Coupled Routing and Excess Storage distributed hydrologic model and the Sacramento Soil Moisture Accounting are both adapted for use in EF5. EF5 is then used to simulate all time series gauged basins in the CONUS with basin areas less than 1,000 km2. The water balance models are then evaluated in terms of bias, correlation coefficient and Nash Sutcliffe Efficiency. The results show that the water balance models have skill over most of the CONUS with the exception for the mountain west where low quality precipitation estimates may be to blame. Finally, a climatology of simulated flash floods is produced over the CONUS by running EF5 to produce gridded daily maximum discharge, time of maximum discharge, and minimum soil moisture outputs. Thresholds are then developed to relate minor flood conditions to basin area and mean annual precipitation so that flooding conditions can be defined even for ungauged watersheds. Maps of the mean annual number of flash flood days are created which show an enhanced region over the central plains particularly Texas and Missouri. Advisors/Committee Members: Gourley, Jonathan (advisor), Palmer, Robert (committee member), Chilson, Phillip (committee member), Kolar, Randall (committee member), Basara, Jeffrey (committee member), Hong, Yang (committee member).